Columbia Law School Scholarship Archive
Faculty Scholarship Faculty Publications
2008
Reason, Reasons and Normativity
Joseph Raz Columbia Law School, [email protected]
Follow this and additional works at: https://scholarship.law.columbia.edu/faculty_scholarship
Part of the Jurisprudence Commons, and the Law and Philosophy Commons
Recommended Citation Joseph Raz, Reason, Reasons and Normativity, OXFORD STUDIES IN METAETHICS, RUSS SHAFER-LANDAU, ED., VOL. 5, 2010 (2008). Available at: https://scholarship.law.columbia.edu/faculty_scholarship/1496
This Working Paper is brought to you for free and open access by the Faculty Publications at Scholarship Archive. It has been accepted for inclusion in Faculty Scholarship by an authorized administrator of Scholarship Archive. For more information, please contact [email protected]. 1 Reason, Reasons and Normativity
Joseph Raz
All normative phenomena are normative in as much as, and because, they provide reasons or are partly constituted by reasons. This makes the concept of a reason key to an understanding of normativity. Believing that, I will here present some thoughts about the connection between reasons and Reason and between Reason and normativity.
1. REASONS AND REASON
Why are the facts which constitute reasons reasons? What about those facts makes them reasons? As expected the answer is that there is an inherent relation between reasons and Reason—understood as our rational faculties or abilities. But what is it? Think of it this way (and I will focus here on practical reasons): that I am hungry is a reason to eat. But what has Reason to do with it? Would I not eat, if hungry and if food were available, anyway? Do not animals of species which do not possess Reason eat when hungry? If reasons do not call for Reason why are they reasons? The facts that are reasons are reasons because they are part of the case for a certain response, for a belief or an action, or an emotion. The example of hunger shows that there can be, and sometimes there are, capacities and processes which reliably lead to appropriate responses to reasons, without the mediation of rational powers. Nor are these absent
Earlier versions of the paper were presented at a workshop at Leeds, at the BSET annual conference of 2006, at CUNY’s Graduate Center in the same year, as well as at a seminar conducted by Christine Korsgaard at Harvard in 2007, and at a conference in Santa Barbara in 2008. I am grateful to participants on those occasions for their comments and responses.
Electronic copy available at: http://ssrn.com/abstract=1008795 6 Joseph Raz
in humans.¹ So why are reasons reasons if there are processes which align us with them independently of our rational powers? We cannot claim that in all domains reliance on rational capacities is better at bringing us into conformity with reasons than other processes. Often the automatic, Reason- bypassing processes are best. Though there need be no rivalry between them, and often rational capacities regularly interact with Reason-bypassing processes. Of course, some reasons can neither be brought into existence nor recognised except through the use of Reason. This is the case with most reasons which are cultural creations. But even in their case, the explanation why they invite certain responses need not invoke that fact. For example, some epistemic reasons which can only be recognised as reasons by rational creatures can nevertheless be identified as facts calling for belief, because they are evidence for the truth of that belief, and their standing as evidence can be explained without invoking the way that they are identified. So why are they reasons? Because Reason is our general capacity to recognise and respond to reasons. There are other capacities which also do that. But Reason is the universal capacity to recognise reasons, one which in principle enables us to recognise any reason which applies to us, and to respond to it appropriately. A little later I will say more about the way Reason enables us to recognise reasons. In particular that it enables people reflectively to recognise that the facts that are reasons are reasons. Here I will add just one observation: that Reason is a capacity reflectively to recognise reasons does not entail that every exercise of reason involves reflection, reasoning or deliberation. With experience we learn to identify and respond to reasons instinctively, though in ways which depend on and presuppose first, reliance on past reflection, and second, the monitoring presence of rational powers which control and stand ready to correct misidentifications or misdirected responses. Even so the statement that Reason is the general capacity reflectively to recognise and respond to reasons may appear formal and uninformative. Is not saying that Reason is the universal capacity to identify and respond to reasons like saying that we dream dreams? Since ‘dreams’ are defined as the objects of dreaming, saying that we dream dreams is a mere formal statement. But the analogy with Reason and reasons ignores the normative
¹ The point is probably limited to practical reasons and reasons for emotions, and does not apply to epistemic ones. The reason is conceptual:ACTIONisagenericconcept,covering acting for reasons, as well as other actions, whereas BELIEF is the specific concept which marks one of various broadly cognitive states which essentially involve our rational capacities, being essentially responsive to reasons (not necessarily in being generated by the realisation that there are reasons for them, but rather in being modified upon awareness that there are sufficient reasons against them), as some of the others, like guesses and hunches, are not.
Electronic copy available at: http://ssrn.com/abstract=1008795 Reason, Reasons and Normativity 7 aspect of the relations between them. Reason can malfunction. Therefore, reasons cannot be defined as what Reason recognises and responds to. As Reason may fail there are criteria by which success and failure are determined, and they determine what reasons there are to be recognised. In conclusion we can say that Reason does not make reasons into reasons (Reason is not a source of reasons). But they are reasons because rational creatures can recognise and respond to them with the use of Reason. Needless to say, there are features of the world that we respond to, where the response cannot be secured via using our rational powers. Such responses are involuntary, and the triggering features are classified as stimuli. Because the response cannot be secured by our rational powers, that is because it cannot be guided reflectively, those stimuli are not reasons, even if the response is sensible, beneficial, etc. To be a reason a fact must be one that we can respond to using our rational powers, whether or not on any specific occasion that is how we do respond to it. Being the general reflective capacity to recognise reasons distinguishes Reason from other processes like hunger, or the instinctive avoidance of fire, which recognise some specific kinds of reasons.
2. REASON AND REASONING
The thesis that Reason is the power to recognise reasons will be finessed and somewhat modified in Section 3. First, I will relate it to the more common view that Reason is the power of reasoning. Paul Grice combines both views: No less intuitive than the idea of thinking of reason as the faculty which equips us to recognize and operate with reasons is the idea of thinking of it as the faculty which empowers us to engage in reasoning.² Grice thought that the two ideas harmonise. He proceeds to explain: Indeed if reasoning should be characterisable as the occurrence or production of a chain of inferences, and if such chains consist in (sequentially) arriving at conclusions which are derivable from some initial set of premises, and for the acceptance of which, therefore, these premises are, or are thought to be, reasons, the connection between these two ideas is not accidental.³ I will basically follow Grice, though my understanding of the way the two ideas harmonise is somewhat different from his. Reasoning is Reason’s main way of recognising reasons. But Reason includes more than the power
² Paul Grice, Aspects of Reason (Oxford: OUP, 2001), 5. ³ Ibid.
Electronic copy available at: http://ssrn.com/abstract=1008795 8 Joseph Raz to reason, and since not all reasoning aims at identifying or operating with reasons the relationship of Reason and the power to reason is more complicated than may at first appear. I have already noted that some mental activities which depend on past reasoning are manifestations of our rational powers even though they do not involve present reasoning, and that activities guided and monitored by Reason in the background are also manifestations of rational powers not involving reasoning. But there is much more. Sometimes it is irrational for people to fail to engage in reasoning, as when they have an overriding reason to reason, and are in a position to know that. Such irrationalities are non-derivative. Even when, as is only sometimes the case, failure to deliberate is a result of a decision, the irrationality of failure to deliberate does not depend on such prior decisions, and does not derive from their irrationality. Likewise, sometimes the very activity of reasoning, even when one is reasoning flawlessly, is irrational. There may be conclusive reasons, which are known or should be known,⁴ not to reason, thus rendering one’s reasoning irrational. One can also be non-derivatively irrational in continuing deliberation for too long, failing to come to a conclusion. Besides, weak-willed intentions show that intentions can be non-derivatively irrational. They are intentions one forms against one’s own better judgement. It follows that the irrationality of a weak-willed intention does not derive from failure to reason correctly. These considerations suggest that Reason consists of more than the power of reasoning, and includes at least the power to form intentions and decisions. Two considerations support this view. First, reasoning being an intentional activity, the power to form intentions and decisions is intimately involved in it, and given that their faulty use can render the whole activity of reasoning on a particular occasion irrational there is a case for counting the power to form intentions and decisions among our rational powers. Secondly, and more generally, I have been implicitly relying on a test which Iwillcalltheirrationality test. It says that if the exercise of a capacity can be non-derivatively irrational (that is irrational not because something else is irrational) then the capacity is one of our rational powers. I will further consider the test in the next section. Still, the ability to reason is at the core of our rational capacities. How so? Some, like Harman, think that reasoning ‘is a process of modifying antecedent beliefs and intentions’.⁵ Butthatshouldnotbetakentoprovide
⁴ Or at least one should accept the proposition that there are such reasons and behave accordingly. ⁵ ‘Practical Reasoning’, in Reasoning, Meaning and Mind (Oxford: OUP, 1999), 46. His later paper ‘Rationality’ (ibid.) allows for reasoning which does not change judgement. Reason, Reasons and Normativity 9 a test for identifying reasoning. It is doubtful whether reasoning has to lead to modification of belief or intention. I may, for example, examine one of my beliefs, and, not having new information, I consider again the considerations I considered before and do not find grounds to change my mind.⁶ My reflections seem a straightforward case of reasoning, but they need not involve a change of belief or intention. There are other cases in which one reasons without changing beliefs or intentions. The reasoning may have been tentative, not reaching any final conclusions. In such cases the reasoning, though not leading to their change, is undertaken in order to examine the case for a modification of one’s beliefs or intentions. This is commonly the case, but it need not be. One may indulge in reasoning in order to pass the time. One may playfully examine hypotheses and their consequences, doing so as a game, a pastime, possibly doing so carelessly, off-handedly. It does not matter. Nor does it matter if one’s reasoning is affected by wishful thinking. One would not change one’s views as a result, for one does not take the activity, the reasoning, seriously. It is just an amusing pastime. One is not irrational on such occasions, even though the reasoning is faulty, and may display a propensity to commit fallacies, which may give grounds to believe that one would be irrational when reasoning ‘seriously’. Bad reasoning, I conclude, is⁷ irrational when and because one non- accidentally fails to respond appropriately to reasons, and the failure is, is due to, a failure or malfunction of one’s rational powers. When one has reasons to form beliefs or to consider the merits of beliefs or actions, then if one does so irresponsibly, carelessly or negligently, or if one’s reasoning is affected by motives which should not affect it, that is if one is guilty of one or another form of motivated irrationality, and if these failures are due to an entrenched disposition to fail in that way, which afflicts one’s rational powers, then one is failing to respond properly to those reasons in a way which renders one’s flawed reasoning irrational. Motivated irrationality is a form of belief-formation contrary to reason and so is negligent reasoning. Negligence is not merely carelessness. It is carelessness (for which we are responsible) when we have reason to be careful. We can only reason negligently when we have reason to form beliefs or seriously to consider the merits of beliefs or of actions. The examples support the conclusion that
⁶ Nor need such cases yield a belief that the beliefs or intentions one examined need no revision, at least it need not yield a belief to that effect different from the one one had before. It depends how much one trusts the reasoning just concluded. ⁷ Perhaps I should say ‘deserves’ to be regarded as irrational—one may well claim that the boundaries of the concept of ‘irrationality’ are vague on this point. It is the theoretical account of Reason which ultimately drives this view. So at most one can claim that it is not at odds with firm features of usage. 10 Joseph Raz mere bad reasoning does not constitute irrationality. I will return to the point below.⁸ This leads to the main point: while not all reasoning aims at identifying and operating with reasons, the power of reasoning is essentially a power whose purpose is to identify and respond to reasons. Reasoning is an intentional mental activity, and a norm-guided activity in that it is governed by criteria of correctness. This is true of both ‘serious’ and ‘non-serious’ reasoning. Reasoning playfully is reasoning governed by the same norms, and mistakes remain mistakes (though their commission does not show that we act irrationally). The norms of correct reasoning show that the point of reasoning is to enable us to detect and respond to reasons. The constitutive standards of reasoning determine both what is reasoning and what is successful reasoning. Reasoning is an activity which is held responsible to those standards, an activity whose success is judged by them.⁹ They determine the success of reasoning as a generic activity, though the actual reasoning one engages in on this occasion or that may be undertaken for some other purpose, and may be successful—though not as reasoning—for some other reason even when flawed as a piece of reasoning. This, then, is the difference between ‘serious’ and ‘non-serious’ reasoning. Serious reasoning is meant to serve its point, and, therefore, to be successful according to its constitutive standards. Non-serious reasoning is the use of the same capacity, performing an activity of the same type, but detached from its normal purpose. The activity is still held responsible to its constitutive standards, but the reason which led to engaging in it may be served regardless of its success as reasoning. So non-serious reasoning is a marginal case of reasoning. It is understood and is broadly conducted as serious reasoning, but, as it is detached from its point, failure to conform to the norms governing the activity is not necessarily, on that occasion, a fault. Reasoning’s constitutive standards are ones which ensure that, when followed, the conclusion of the reasoning is warranted, namely one which
⁸ Note that there are two kinds of mistake in reasoning which do not render it irrational. In the text I focus on one of them, i.e. when one is not reasoning ‘seriously’, one is not reasoning in response to a perceived adequate reason to reason. The other is when the mistake is due to a lapse of either memory, attention, or another of one’s ancillary powers, which are not among the rational powers, or to an occasional rather than entrenched malfunction of one’s rational powers. ⁹ This gives rise to the question: when is an activity governed by those standards? I will not consider this issue here. It seems reasonable to suppose that it is so when either the person engaging in the activity is one who has the capacity to reason and takes himself to be so governed, namely accepts the legitimacy of judging his activity by those standards, or if that person behaves in a way which is reasonably understood as presenting himself to others as someone who is reasoning. But much more needs to be said. Reason, Reasons and Normativity 11 we have adequate reason (though not always a conclusive reason) to accept. Directly they warrant belief, indirectly they warrant actions, emotions and other reason-sensitive attitudes. To be more accurate we need to take note of the fact that in most instances of reasoning one relies on some propositions whose credentials are not examined during that reasoning episode. Hence successful reasoning assures reasoners that the conclusion is warranted, on the assumption that so are the unexamined propositions relied upon. Its purpose is to establish that given that assumption, certain beliefs, intentions, and the like are warranted. The norms of correct reasoning determine the point of reasoning, and since the norms of correct reasoning are norms that warrant acceptance of the conclusion of a correct reasoning the purpose that they serve is to guide us in judging which beliefs and intentions we have reason to have. This establishes the connection between Reason as the capacity to respond to reasons, and reasoning.
3. THE SCOPE OF REASON
Characterising Reason as the general reflective capacity to recognise reasons raises the question of which mental capacities belong with it, which belong with our rational powers? I will say little on the subject. It is not my aim to suggest criteria capable of adjudicating various borderline cases, either between animal species which typically have Reason and those which do not, or regarding the boundaries within a species between those who have Reason and those lacking it. The fate of borderline cases may well depend on additional considerations not canvassed here. But a few further observations on the core concept, relying on our general knowledge of it in a way which ties it to the account here proposed, may be helpful. Various connections between Reason and other concepts can be called upon in clarifying the concept. One is between Reason and personhood: only persons have Reason. A second is between Reason and accountability, which marks one sense of responsibility. Creatures that do not have rational powers are not responsible (accountable) for their actions.¹⁰ And a third is with the notion of irrationality. It led to the irrationality test suggested above. The irrationality test suggests demarcations of Reason which are not dictated by its characterisation as a general power to recognise and operate
¹⁰ This does not mean that we are only responsible for or accountable for the exercise of our rational powers. See my ‘Responsibility & the Negligence Standard’, Oxford Journal of Legal Studies 30(1) (2010), 1. 12 Joseph Raz with reasons. Decline in some mental capacities like the powers of memory and of concentration affects one’s ability to detect and respond to reasons. But we do not stop being persons simply because of loss of memory and of concentration (at least not while we retain them to a minimal degree). Hence these powers do not belong with our rational powers. But if so, given that our success in recognizing reasons depends on our powers of concentration and memory, can Reason be identified with the reflective power to recognize reasons? This identification of Reason is correct, so long as it is understood to represent the core of the concept of Reason, and is not expected to set its limits. The irrationality test is one of a number which offer more help with marking the limits of our rational powers. It too suggests that memory and the power of concentration are not among them because failures of memory and concentration, however bad, are not irrational. They are among many mental capacities which are ancillary to our rational powers, enabling them to function well. Some provide the input on which the rational capacities rely through perception, sensations (including those which are accompanied by drives: like hunger, thirst, discomfort at the surrounding temperature, sexual arousal, and more), or memory. By the irrationality test the powers of reflection, deliberation and decision are rational capacities. Interestingly the irrationality test suggests that the capacity to have emotions is among our rational capacities because emotions can be non-derivatively irrational. The test relies on a distinction between derivative and non-derivative irrationality. Something is derivatively irrational if it is irrational only if and because something else is irrational. For example, actions are irrational only if and because the intentions or decisions which render them intentional are irrational.¹¹ Cases of non-derivative irrationality are cases where the irrationality of what is irrational does not logically or conceptually depend on nor is it due to some other irrationality. If a flawed exercise of a capacity can only be derivatively irrational that capacity is not part of Reason. If, on the other hand, a flawed exercise of a capacity can be non-derivatively irrational then that capacity belongs with our rational capacities, is part of our Reason. Arguably emotions can be irrational because they are founded on irrational beliefs. Fear based on an irrational belief that the shadow on the wall is the devil about to kill one is an irrational fear. But emotions can be irrational when they are disproportionate reactions to rational beliefs. In such cases their irrationality is non-derivative, and
¹¹ One interesting result is that only intentional actions can be irrational, and not even all of them, since not all intentional actions are actions undertaken with an intention. Reason, Reasons and Normativity 13 therefore by the irrationality test, emotional capacity is part of our rational powers. This conclusion will not surprise some, while seeming preposterous to others, who will take it as a reason to reject the irrationality test. There is, however, an independent case for including the capacity to have emotions among the rational powers. For example, empathy is crucial for understanding other people, as well as animals of other species, and arguably emotional responses are essential to our ability to understand that we have reasons of certain kinds, and to the ability to understand¹² what response to various reasons is appropriate, as well as to motivate us to respond as we should. There are also separate reasons for confidence in the irrationality test. It seems natural to think that only failures of the power of Reason could be irrational, except when the irrationality is derivative, that is, when it derives from failure of powers of Reason. However, these matters require a more detailed exploration of the role of the emotions in our make-up as persons, in our motivations, and in our cognitive powers. There is one important clarification, indeed modification, of the slogan that Reason is the general reflective capacity to recognise reasons. The slogan may give an unduly passive image of rational powers, just tabling reports, as it were, of what reasons are to be found where. The slogan should be augmented to say that Reason is the general capacity to recognise and respond to reasons. Clearly that is so regarding theoretical reasons: recognising a sufficient case for a belief is adopting the belief. There is no separate step involved, no transition which, pathological cases apart,¹³ can fail. Properly recognising epistemic reasons is properly responding to them. Things are somewhat different when it comes to practical reasons (including reasons for mental acts). Action may require interventions in the world, regarding whose success agents have less control than over the response to epistemic reasons or to reasons for mental acts, intentions, and omissions (on most occasions). Except for the capacity to reason, decide, form intentions and a few other capacities essential to be able to act with intention which are part of our rational powers according to the irrationality test, the capacity to act is not part of our rational capacities. This has to be borne in mind in interpreting the extended slogan.¹⁴
¹² We can learn of those reasons, and of the response they make appropriate, from testimony or other indirect means. But arguably we can understand them only because we have emotional capacity ourselves. ¹³ Psychological phenomena are ‘pathological’ when they meet some of the criteria determining their attribution (or existence), while failing others. ¹⁴ A capacity to act, while not part of the rational capacities, is an essential part of the wider group which constitute the capacity for rational agency. They are pivotal in determining 14 Joseph Raz
Nevertheless the extended slogan, properly understood, is correct. For even regarding practical reasons, rational capacities must involve response to them if they are to involve recognising them. People who recognise a conclusive reason to (to eat, or whatever) and who fail to respond to it at all, fail (when the time comes) to form an intention to ,haveno positive attitude at all towards -ing, do not respond appropriately to other people -ing, etc., are non-derivatively irrational. Thus, the irrationality test shows that capacity properly to respond to reason is part of our rational capacities. Furthermore, people who fail to respond appropriately in any way at all do not fully recognise the existence of the reasons. Attribution of belief depends on the existence of a variety of criteria of belief, and they include not only avowing the belief, and attesting to reasons for it, etc., but also responding to it appropriately: those who would not put an apple on a table (assuming normal circumstances), for fear that the apple may fall to the floor, show themselves not to believe that there is a table there. At best theirs is a pathological case of belief. Hence recognising reasons involves responding to them, and the mental capacities involved in setting ourselves to respond, the powers of decision and intention, are part of our rational powers.
4. IRRATIONALITY
Reason, i.e. the rational powers or capacities, is involved in activities such as choosing, deciding, reasoning. These activities, and therefore their results,¹⁵ are rational so long as the rational powers guiding them function properly. They are irrational when these powers malfunction. It is possible, by accident, for the result to be a happy one, that is the person concerned may accidentally choose or decide as reason directs. But when the result is the outcome of a malfunctioning rational power it is happy but irrational. This line of thinking allows for a distinction between mistakes (in reasoning, choosing, etc.) which render the activity irrational, and those which do not. The latter are due to malfunctioning of powers relied on in these activities which are not among the rational powers. For example, responsibility. See ‘Responsibility & the Negligence Standard’. And ‘Being in the World’ (forthcoming), ‘Agents and their Actions’, special issue of Ratio. ¹⁵ Using the term in the technical sense given it by von Wright (Norm and Action (London: Routledge and Kegan Paul, 1963)) i.e. as the state which defines the completion of the process or activity. Reason, Reasons and Normativity 15 they may be due to lapses of memory, failures of attention, or failure of perception, etc. Possibly, the account has to be modified somewhat to allow for additional kinds of mistakes which do not render the activities irrational. (For example, cases where the malfunction is accidental may not render the activity and its outcome irrational. Arguably, they are irrational only if due to a persistent condition.) I will not examine that possibility here. This view of rationality—the view that we are rational so long as we are properly guided by well-functioning rational powers—has been followed by various writers, e.g. Michael Bratman. Others, e.g. John Broome, have taken a different view: according to them rationality relates to the occurrence or non-occurrence of certain relations between mental states. For example, and to simplify, our beliefs are irrational if they are contradictory.¹⁶ By way of abbreviation I will call the first view dynamic and the second static. How do the two views relate to each other? They would be in harmony if it turns out that they coincide in their judgements about the rationality of beliefs, intentions, emotions, etc. But they do not. We may reach conditions condemned as irrational by the static view by reasoning as we must according to the dynamic view. For example, there may be overwhelming evidence for a certain proposition and once aware of that I believe that proposition. I have no choice, and I am dynamically rational. But it is possible that I am also aware that the proposition contradicts some of my other beliefs. I am therefore aware that either this or some other of my beliefs are false, but I dynamically-rationally think that the fault is in the others. I do not know which of the others is the guilty one. I only know that the proposition concerned contradicts one or another of my other beliefs. Nor do I know how to improve my other beliefs (one of them is false, but it may approximate the truth). This may occur in conditions in which it would be dynamically irrational for me to abandon any of my current beliefs, but dynamically irrational not to add to them a belief which, as I know, is inconsistent with them. If situations like this are possible then the two approaches to rationality conflict. They also conflict because one can satisfy the static conception without satisfying the dynamic conception (e.g. via a string of mindless accidents).¹⁷ It is possible to say that there is no conflict for these are not two rival accounts of rationality. Rather they are explanations of two different concepts of rationality. I cannot examine this possibility
¹⁶ Broome lists a number of more carefully formulated relations between mental states whose occurrence renders them irrational. ¹⁷ A point made to me by Niko Kolodny, whose letter to me I am here quoting. 16 Joseph Raz here, so let me simply record that I suspect it to be mistaken, i.e. I suspect that the two accounts do conflict, and that at least one of them is mistaken.
5. NORMATIVITY AND PERSONHOOD
Finally, I turn from the relations between Reason and reasons, to the relations between Reason and normativity. First, I will reject two tendencies sometimes found in discussions of normativity. One is to identify it with an orientation towards value, the other with the demands of ration- ality. I will suggest that the value of a thing provides some reasons, but not all, and in any case does not explain the force of reasons. I will also suggest that there is no reason, and no need for a reason, to be rational.¹⁸ The difference between practical and epistemic reasons is central to the attempt to understand the normativity of reasons. It defeats any attempt to explain normativity as having to do with the influence of value on us. Epistemic reasons have nothing to do with value. One way to persist with the view that normativity is tied to value is to claim that reasons for belief are only conditionally normative, which means that they are only conditionally reasons. They are reasons when it is good to have that belief, but not otherwise. Epistemic reasons help towards having true beliefs, and this matters only when there is a reason to have them. Epistemic reasons are a type of instrumental practical reasons. This view flies in the face of the fact that all judgements which go with reasons apply to epistemic reasons unconditionally. To mention but one: we are not irrational for failing to conform to a conditional reason until the condition is met. But it is irrational not to have the beliefs one has adequate reason for whether or not the condition obtains, that is whether or not there is value in having them. Next, a few words about Rationality and normativity. First: neither the question of the hold epistemic reasons have on us nor the question of their normativity is the question ‘why be rational?’ We can fail to conform to reasons which apply to us and still be rational. That would be the case so long as our failure is not due to a malfunction of our rational powers, for example, so long as our failure is due to non-culpable (e.g. non-negligent) mistakes and ignorance. There is no normative standing to being rational as such. There is no reason to be in a state in which one failed to identify
¹⁸ I have argued for this point at greater length in ‘The Myth of Instrumental Rationality’, Journal of Ethics & Social Philosophy 1(1)(2005),
¹⁹ Of course we do not commend a person for failing non-negligently to identify reasons which apply to him, and therefore we do not say to him ‘you were very rational in what you did’. But that is a point about the implications of what we say. At other times we may say so: ‘Oh, I have become totally irrational’ he exclaims in despair, and we reassure him: ‘You are perfectly rational, you just made a mistake.’ ²⁰ In a comment on an earlier version Kolodny explains the point better than I did: ‘True, my believing that 2+2=4, knowing what I know, is not (i) under my voluntary control. Nevertheless, my believing it can be (ii) the direct upshot of deliberation, of reflection on reasons (and this amounts to a kind of control over my beliefs, . . . ). By contrast, not even the latter is true of the functioning of my rational powers as a whole. There is no question of deliberating, ‘‘Shall I function rationally?’’ and then directly proceeding, on the basis of an affirmative or negative answer, to continue functioning rationally, or to cease to. I could not follow such a reason, in the sense Raz emphasizes in ‘‘Reasons: Practical and Adaptive’’ (in D. Sobel & S. Wall, Reasons for Action, (CUP 2009) 37).’ 18 Joseph Raz epistemic reason, what is wrong with that? The answers seem to be internal to the concepts involved: we say—if you defy reason you are irrational. So what? Your beliefs are incoherent—well suppose they are, what of it? And so on. The temptation to say: ‘if you disregard reasons you will fall down and break you neck’, or point to some other adverse practical result, is overwhelming. And as we saw has to be resisted. We, some of us, want to step outside the conceptual web, and find an explanation for the hold that reasons have on us. This we can do since their hold on us depends on the fact that responsiveness to reasons is constitutive of personhood. I should immediately make clear that I do not share the thought of some philosophers that a constitutive account of reasons will settle what reasons there are. Thinking of practical reasons, Korsgaard, e.g. claims that ‘Action isself-constitutionandaccordingly,...whatmakesactionsgoodorbadis how well they constitute you.’²¹ She suggests that all practical reasons can be derived from this insight: we have a standard of being a good person, and you have reasons for actions which will constitute you as one, or something which includes this idea. My constitutive account of rationality can yield no such results. It is a mere formal account, and we will have to consider whether such a formal account is, even if true, of any significance.²² Here is the formal constitutive story: reasoning and deliberation are mental activities which we can decide on. There are practical reasons for or against reasoning and deliberating. But we cannot decide how to respond to epistemic reasons. It is constitutive of belief that it is governed by our responsiveness to epistemic reasons, governed gaplessly, automatically as it were. Responsiveness to epistemic reasons is constitutive of believing. Responsiveness to reasons can, of course, fail: through mistakes, fallacies, wishful thinking, self-deception and more. How then are these standards constitutive of belief? Does not failure establish that conformity to reason is only contingently related to belief? Conformity to reason is indeed contingent. What is essential to belief is, first, its subjection to the normativity of reasons, its being subject to evaluation as warranted or unwarranted depending on its conformity with reasons; and, second, the fact that it is automatically, as it were, self-correcting. Failures to conform to reasons are self-correcting when we become aware of them. Again, no gap exists, no decision to correct is required, no involvement of the will.
²¹ C. Korsgaard, Self-Constitution: Action, Identity and Integrity (Oxford: OUP, 2009), 25. ²² This makes it immune to the criticism of constitutive accounts advanced by David Enoch in ‘Agency, Shmagency: Why Normativity Won’t Come from What is Constitutive of Agency’, Philosophical Review 115 (2006), 169–98. Reason, Reasons and Normativity 19
That is why the responsiveness to epistemic reasons is a form of con- stitutive normativity, normativity built into the very possibility of belief.²³ Not all our mental states are responsive to reasons. I may think that I am a feathered bird, and not respond at all to evidence that I have no feathers nor beak, etc. But then that kind of thinking is fantasising, imagining, day-dreaming or the like; not believing. There are two lessons here: first that we recognise the difference ourselves. We know the difference between belief and imagination, and we know that it consists in part in that belief is, while imagination is not, subject to the full discipline of reasons (though imagination may be partially subject to it, in a variety of different ways). I say that we know that, meaning not that we would, or even could, articulate what we know in these terms. We may not be that reflective, we may have false philosophical beliefs, or we may lack the concepts I am using. What I mean is that the statement I made correctly describes what we know, regardless of whether or not we are aware of this. The second lesson is that our control over belief differs from our control over various forms of imagination. Regarding the imagination control consists entirely of voluntary control. It depends on the degree to which we can imagine at will. Not so with belief. We cannot always imagine things at will, and when we do we cannot always make ourselves stop imagining at will, and, most noticeably, even when we imagine something at will our voluntary control over the details of the imagining is very limited. In all these respects our voluntary control over belief is generally similar (we can decide what to deliberate about, have some ability to remember, i.e. to recall, at will, can often decide to stop considering a matter, and so on). But distinctively, we have little voluntary control over the way our beliefs respond to reasons we think we have. We also have limited voluntary control over whether to have beliefs. We can shut our eyes, but when open we cannot just refuse to believe what we see
²³ Naturally, the full story is much more complex. One point in particular is crucial for the coherence of my account: some conditions of rationality, i.e. some conditions of responsiveness to epistemic reasons, are universally constitutive of belief, so that there is no (non-pathological) belief which does not conform to them. Some of these apply when agents are aware of their application. An example of these may be the condition that one cannot, in full awareness of the fact, believe a proposition and its negation (though one can oscillate between them, or display other pathologies). We are sensitive to some conditions of rationality even when we are not aware of them as conditions of rationality, in that they tend to increase the degree to which we hold beliefs only if warranted through consciousness-bypassing processes. But generally, firm belief that something is a condition of rationality (whichever way it is conceived, i.e. even if one is not aware of it under this description) will have a more direct effect on our belief formation. That enables us to improve our understanding of rationality, and of the conditions of rationality, and thereby improve the processes by which we form and maintain beliefs. 20 Joseph Raz
(whatever we say) without believing that we have some reason to doubt what we see. What is important is that the limits of our voluntary control over our beliefs are not the limits of our control over them: we are in control over our beliefs by functioning properly as rational agents, that is, we are in control, and active, so long as, and to the degree that, our beliefs are governed by Reason, by our rational powers.²⁴ That is what makes us persons. Roughly speaking we are persons as long as we have rational capacities, and by and large our beliefs and actions are governed by them, which is the same as saying so long as we have beliefs. In brief outline: responsiveness to practical reasons is also constitutive of being a person, for without it there is no action with the intention of doing it. While not all actions are performed for a reason, when we do something with the intention of doing it, which is roughly when we have a purpose in doing it, see a point in doing it, we act for a reason, that is we act in the belief that there is a reason for the action. Such actions, which I will somewhat inaccurately refer to as intentional actions, are governed by reason somewhat less directly than beliefs. Both intentional action and belief are subject to failures correctly to identify reasons, failures to follow through with the implications of reasons one identified (as in failures of memory or lapses of attention) as well as to irrationalities, including motivated irrationalities. In both cases we recognise such failures for what they are in principle, thus acknowledging that beliefs and intentional actions are inherently governed by reason. There is, as we know, a keen debate about the difference between the conditions for rationality of belief and of intentional action, e.g. whether there is epistemic akrasia. For our purpose suffice it to note that whatever differences there are they are secondary to the basic point that we cannot, while acting intentionally, but act for a (perceived) reason (albeit not always the one we believe to be the best reason).²⁵ Rationality, namely responsiveness to reasons, is thus constitutive of being persons. The relationship of Reason and reasons has been the thread going through this paper. Trying to understand what motivates one to search for, say, the foundations of normativity, we realise that the search cannot even as much as be stated from ‘inside’ as it were. From ‘inside’ we can only look for reasons for reasons, or note that disregarding reasons is, under appropriate circumstances, irrational, that is a symptom of the malfunctioning of our
²⁴ Compare my discussion of the matter in Engaging Reason, ch. 1 (Oxford: OUP, 1999). ²⁵ I have discussed challenges and exceptions to this view on various occasions, most comprehensively in chapter 2 of Engaging Reason, and in ‘The Guise of the Good’, in S. Tenenbaum (ed.), Desire, Good, and Practical Reason (Oxford: OUP, 2010). Reason, Reasons and Normativity 21 rational powers. All such internal investigations inevitably move in a circle, and do not reach the puzzle. So we went outside, and raised a question about reasons, stated in non-normative terms: what is the hold reasons have on us? The answer wasthatwecannotignorethembecausewearepersons,ormoreprecisely, because rational powers are constitutive of personhood, and because they are powers whose use does not depend on our will. That is, these powers are engaged and active willy nilly, independently of any decision to use them, so long as we are awake and do not suppress them. They are like our hearing rather than like our ability to speak. Hence, so long as we are persons we engage with reasons, generally trying to do it well, however imperfect our success. Now, you may object that even if true these observations do not explain normativity, let alone justify its standing. They say something about rational powers, but nothing at all about reasons. I emphasised early on that explaining why reasons call for a certain response need not invoke our rational powers, which are merely powers to recognise and respond to what is there independently of them—well, at least sometimes, given that some reasons would not exist but for the powers and activities of rational creatures. In a way the objection is justified. Rejecting the feasibility of a reductive explanation means that once we step outside, as it were, and examine normativity as a whole we lose the ability to explain it. That explanation is inevitably internal—reasons are what we should follow, disregarding them is unjustified, etc. etc. But we can explain from outside the inescapability of normativity, the hold reasons have on us. We do so, to be sure, by pointing out features of rational powers, but rational powers are essentially powers to recognise and respond to reasons. So in explaining their place in our life we also point to the hold that reasons have on us, though that hold is subject to mistakes and irrationalities. Still, disquiet may persist: does any of this amount to more than saying that we are boxed into treating certain considerations as if they were reasons? Does it not fail to vindicate them as reasons? Are reasons normative in the deeper sense that there is a point to being guided by them, that being guided by them is not an arbitrary, albeit a natural response? Is our responsiveness to reasons just a fact rather than a response to a normative consideration? Is it not like saying that we are persons only so long as we breathe, a fact which we must acknowledge, but can find arbitrary, and are free to resent? I’ll say nothing about breathing. But reasons are different. Here, I think, the fact that the account is confined to their formal characterisations is—far from being a shortcoming—a key to its success. Reasons are governed by maxims stating that a reason for belief is a fact which is part of a case for 22 Joseph Raz having that belief ,anda reason for action is a fact which is part of the case for performing that action. Such characterisations are relatively formal in that they do not directly yield a way of establishing what is a reason for what. But they focus discussion about what properties make for reasons: such investigation may yield that beliefs, though not thoughts generally, are without flaw if true, and actions done with intentions to do them, though not actions generally, are without flaw if they are adequately valuable. Hence reasons for beliefs are truth indicators, and reasons for action are their value properties. Truth indicators are truth indicators even if there are no rational creatures who can rely on them to form beliefs. Values, those whose existence does not depend on the culture of rational creatures, are values, and have instances whether or not there are persons able to perceive and respond to them as values. Our rational capacities enable us to recognise and respond to reason-constituting facts, and being reflective powers they enable us to improve our understanding of what makes those facts what they are, and how best to identify them for what they are. Do all values derive from the well-being of some agent? Is there some property which constitutes some facts as moral reasons, and so on? The formal characterisations serve to focus and frame thought about which more substantive properties constitute reasons, properties whose standing does not depend on our rationality. This, some will say, is a very surprising view to end up with. Does it not mean that normativity resides in our rationality; that values, and truth indicators, etc. are not normative in themselves? It depends on what you mean by in themselves. Barometric pressure is evidence of weather which we, using our rational capacities, have come to recognize and we form beliefs on its basis. No doubt some other animal species, lacking rational powers, respond ‘automatically’ to variations in barometric pressure. They do not take it to be a reason for anything. They do not recognize its normative aspect. But it has it. It has it in virtue of the ability of rational creatures to recognize that it is evidence of forthcoming weather. Or think of our dependence on food for survival. We forage for food using our rational capacities. Some other animal species seek food independently of any rational powers. It is no reason for them, but it is of value to them. We depend for survival on our ability to adjust our pupils to light levels, our ability to adjust our breathing and blood circulation to prevailing conditions. We do so as other species react to barometric pressure: the features are not reasons for us. But they are good for us. Does all this show thatvalueandevidencearenotnormative?Atmostitshowsnotthatthey are not normative, but that what makes them normative is their ability to function as reasons, that is to be recognised as such by creatures using their rational powers. Reason, Reasons and Normativity 23
But, some will say, the question remains: assume that I am right in emphasising (a) that subjection to the discipline of reasons is semi- automatic, (b) that it is normative in being an adjustment of our beliefs and intentions in light of reasons, and (c) that the capacities manifested in these adjustments are constitutive of being persons. But having conceded that this is so only shows that a general rational capacity is constitutive of personhood, and not that reasons or normativity are constitutive of personhood: the three points fail to vindicate normativity. It is not enough to point out that creatures with no rational capacities are not persons. Clearly there is nothing amiss with pigeons, even though they do not (I assume) have rational capacities. A vindication of normativity has to show what is amiss with failing to conform to reason on this or that occasion. At this point I have to admit that I no longer understand the sense of puzzlement. What is amiss with failing to conform to reason is just that. It can be specified further: it may be defrauding a person of his money, or it may be wasting one’s talent, or missing an opportunity to make a lot of money, or remaining confused about black holes. It all depends on the nature of the reasons one flouted. But clearly that is not the puzzle. It has something to do with vindicating reasons or normativity in general, without assuming their cogency. So what is it? We do know that people who flout reason sometimes prosper. Is the desire for some further vindication of reasons a hope that philosophical argument can show this to be an illusion? But there is no illusion there.